Reading comprehension difficulties in children with ... · [email protected] or...
Transcript of Reading comprehension difficulties in children with ... · [email protected] or...
1
Reading comprehension difficulties in children with rolandic epilepsy
Nicola K. Currie1, Adina R. Lew
Department of Psychology, Lancaster University, UK
Tom M. Palmer
Department of Mathematics and Statistics, Lancaster University, UK
Helen Basu
Paediatric Neurology Department, Royal Preston Hospital, UK
Christian de Goede
Paediatric Neurology Department, Royal Preston Hospital, UK
Department of Health Research, Lancaster University, UK
Anand Iyer
Department of Neurology, Alder Hey Children’s Hospital, Liverpool, UK
Kate Cain1
Department of Psychology, Lancaster University, UK
Word Count = 3037
1
Corresponding authors:
Nicola K. Currie or Kate Cain, Department of Psychology, Lancaster University,
Lancaster, LA1 4YF, UK, Tel. 00 44 1524 593990, Fax. 00 44 1524 593744,
2
ABSTRACT
Aim Difficulties in reading comprehension can arise from either word reading or
listening comprehension difficulties, or a combination of the two. We sought to
determine whether children with rolandic epilepsy (RE) had poor reading
comprehension relative to controls, and whether such difficulties were associated with
word reading and/or general language comprehension difficulties.
Method In this cross-sectional study, children with RE (n= 25; 16 males; mean age 9
years:1 month, SD 1 year :7 months) and controls (n= 39; 25 males; mean age 9:1, SD
1:3) completed assessments of reading comprehension, listening comprehension,
word/non-word reading, speech articulation and non-verbal IQ.
Results Reading comprehension and word reading were worse in children with RE,
F(1, 61) = 6.89, p = .01, p2
= .10 and F(1, 61) = 6.84, p = .01, p2
= .10, with
listening comprehension being marginal, F(1, 61) = 3.81, p = .06, p2
= .06. Word
reading and listening comprehension made large and independent contributions to
reading comprehension, explaining 70% of the variance.
Interpretation Children with RE may be at risk of reading comprehension
difficulties. Thorough assessment of individual children is required to ascertain
whether the difficulties lie with decoding text, and/or with general comprehension
skills.
Word count: 198
What this paper adds:
Children with RE may be at risk of poor reading comprehension.
For some children, this was related to poor word-reading skills, whereas for
others it was related to poor listening comprehension abilities, or both of these
combined.
Educators and speech therapists need to conduct assessments of word reading
and listening comprehension, in addition to reading comprehension, to tailor
interventions to the profile of difficulties presented by individual children with
RE, when remediating poor reading comprehension.
RUNNING FOOTER: Rolandic epilepsy and reading comprehension
3
Reading comprehension difficulties in children with rolandic epilepsy
Rolandic epilepsy (RE: or benign epilepsy with centrotemporal spikes,
BECTS) is reported to account for 8 to 25% of all cases of childhood epilepsy in the 5
to 14 year range. 1,2
The onset occurs between 3 and 13 years and the condition
resolves by adulthood. 1
Children with RE have an increased risk of reading problems
in school 3
where reading comprehension is essential for learning across the
curriculum.4
More broadly, reading comprehension in school is a critical determinant
of academic and professional success. 5
Two sets of skills enable reading
comprehension: word recognition skills and listening comprehension.4 These can fail
independently, or together, resulting in poor reading comprehension. Poor word
recognition can result in poor reading comprehension because slow or inaccurate
word retrieval processes can lead to an information processing bottleneck.6 In
contrast, specific reading comprehension difficulties have been documented in
children who have age-appropriate word reading but weak listening comprehension. 7
A recent meta-analytic review confirms poor single word reading in children
with RE 1, and language difficulties at the word, sentence and discourse level have
been found 8 in addition to teacher reports of difficulties specifically with reading
comprehension. 9
However, studies reporting standardised measures of discourse-level
reading or listening comprehension for children with RE are lacking. Our primary aim
was to determine if discourse-level reading comprehension is an additional area of
weakness in this population. To do this, we examined reading comprehension in a
group of RE children relative to age and gender matched controls. A second main aim
was to examine the contributions made by word reading and listening comprehension
in the prediction of reading comprehension. It is important to understand the source(s)
of any reading comprehension difficulties associated with RE, to target appropriate
remediation.6
Pal and colleagues have shown that children with RE have an increased risk of
speech sound disorder (SSD), a developmental condition that resolves around 5-6
years 3 and is associated with reading difficulties in non-RE populations.
10 Children
with RE can also show more generalised dyspraxic difficulties 11
which have been
related to literacy difficulties.12
Thus, we measured SSD and Developmental
Coordination Disorder (DCD) in our sample (through parental questionnaire) to
ascertain if any reading difficulties were associated with these comorbidities. We also
measured children’s ability to pronounce different speech sounds to determine the
extent of current articulation difficulties, given the early resolution of SSD.
Children with RE completed assessments of reading comprehension, single
word and non-word reading, and listening comprehension. We also collected
information about their speech and motor skills. The following questions were
explored:
1. What is the level of reading ability in children with RE in comparison to
typically developing controls? Based on previous research, we predicted that children
with RE would do worse on measures of real word and nonword reading, and reading
and listening comprehension.
2. Are reading comprehension outcomes in RE predicted by word reading and
listening comprehension, over and above age and general cognitive ability, to the
same degree as found for typically developing controls, or is one component skill the
more critical determinant, as in populations with predominantly word reading
difficulties or listening comprehension problems (e.g., children with dyslexia or poor
comprehension)?
4
3. Is there any evidence that SSD or DCD are more likely associated with
reading difficulties in children with RE than in controls?
4. Are clinical variables (age of onset, medication status) associated with
reading difficulties?
METHOD
Participants
Children with RE were identified between October 2013 and December 2014
at 11 participating hospital trusts in Northern England. The inclusion criteria were
children aged between 6 and 12 years with at least two observed seizures and
confirmatory EEG, as assessed by a paediatric neurologist (either CdeG, HB or AI).
All were English-speaking. In terms of sample size, there was no prior basis for
estimating the effect size of the reading comprehension difference between RE
children and controls. Therefore using the average effect size of 0.72 reported in a
recent meta-analysis for single word reading1, an alpha level of 5% and power of
80%, a sample size of 32 RE children and 32 controls is indicated.
A research nurse from the Medicines for Children Research Network together
with local site Principal Investigators identified potential cases and made initial
contact. Interested families gave consent to be contacted by the research team.
Following testing (January 2014 to July 2015) the medical notes of the final sample of
children were reviewed by one of three Consultant Paediatric Neurologists (CdeG,
HB, AI) for full clinical data. Four children were excluded at this stage due to
insufficient evidence for rolandic epilepsy (see Figure 1). Our final sample was 25
children.
The control sample was recruited from three mainstream primary schools in
the North West of England and via research study advertisement at Lancaster
University. They had no known neurological or neurodevelopmental conditions, or
diagnosed reading difficulties (based on parental report), and spoke English. From
sixty initial recruits, only those with a birthday within six months of a child with
epilepsy and the same gender and school year group were included in the data
analyses reported here. As a result all RE children had at least one match but some
had several matches, one child did not have a gender match. This gave an overall
control sample of 39. See Table I for participant characteristics.
A National Health Service Research Ethics Committee (North West –
Liverpool East) and a University Research Ethics Committee approved the study.
Parents of the RE and control sample gave written consent and children gave verbal
consent prior to commencement of testing.
Measures
Children completed standardised tests assessing reading and listening
comprehension, word reading, non-verbal IQ and speech articulation. Re-test
reliability (unless otherwise stated) and test validity are referenced for each from the
test manual or papers assessing validity. Parents completed questionnaires on their
child’s language and motor coordination.
The York Assessment of Reading Comprehension 13
(YARC) was completed.
Children read aloud two short stories and answered eight open-ended comprehension
questions after each one. The time taken to read each text, the number of word
reading errors made, and the number of correct responses to the comprehension
questions were used to calculate separate scores for reading rate (reliability = .90 -
.95, depending on age), word reading accuracy (.75 - .93) and reading comprehension
5
(Cronbach’s = .71 - .93), respectively. 13
The test has validity relative to an alternative
standardised measure of comprehension. 14
Children completed measures of sight word efficiency and phonemic decoding
efficiency from the Test of Word Reading Efficiency (TOWRE) Second Edition15
.
The number of correctly read words/non-words in 45 seconds was recorded. This test
has good reliability (.86 to .93) and validity, demonstrated by correlations with other
assessments of word reading.15
The Understanding Spoken Paragraphs subtest of the Clinical Evaluation of
Language Fundamentals (CELF) 4th
edition assessed listening comprehension.16
Children heard three stories and answered open-ended comprehension questions after
each one. Reliability is adequate to good (range of .51 to .87 depending on age).
Validity is good, evidenced by correlation with other subtests and discrimination in
terms of performance by typical and clinical samples.16
Non-verbal IQ was assessed using the matrix reasoning subtest from the
Wechsler Intelligence Scale for Children IV-UK. 17
This subtest has good reliability
(.71 to .85, dependent on age) and validity (correlation of .84 with overall
performance on WISC perceptual reasoning index).17
Parents completed a questionnaire incorporating items used in previous
research 3, 10
to assess the presence of SSD early in development. Children with a
reported history of speech and language problems, who had received speech/language
therapy, and had delayed production of one- and two-word utterances are considered
likely to have experienced this disorder. 3, 10
On this basis, we coded children who met
all of these criteria as possible SSD cases (see Table I). Children’s articulation of
consonant sounds in English was assessed with the Goldman-Fristoe Test of
Articulation (G-F)18
. Reliability is high (.98) and validity has been demonstrated by
comparison with a Canadian English-speaking typical sample. 18
Parents completed a DCD questionnaire, (DCDQ’07)19
with good reliability
(.89, Cronbach’s alpha) and clinical validity.19
Cut-off scores for possible DCD are
provided in relation to age.
Procedure
Children were tested individually at their home, school or the university,
depending upon parental preference. All assessors were experienced testers with PhD
level training. The measures were part of a larger battery of language, literacy and
memory assessments and were administered in the same order for each child, with
appropriate breaks. Twenty children were assessed in several sessions split over the
course of a day; the remainder in sessions spread over several weeks.
Statistical analyses
Given the wide age range being studied we used raw scores in all analyses,
entering age as a predictor or covariate where appropriate. To examine group
differences, we conducted Analysis of Covariance (ANCOVA) for the measures of
reading comprehension (YARC), listening comprehension (CELF) and non-verbal IQ
(WISC), with age entered as a covariate. All assumptions of the test were met (normal
distribution of measures and homogeneity of variances). For word reading, we formed
a composite score averaging word reading accuracy and reading rate from the YARC,
and word and non-word reading from the TOWRE-2, as these were highly inter-
correlated (rs =.87-.95). An ANCOVA with age as a covariate was conducted on this
score. Fisher’s exact tests were conducted to compare the number of children in the
RE and control sample with significant impairment on the measures (excluding non-
6
verbal IQ and listening comprehension, which are designed to be used as part of a
composite).
To determine whether word reading and listening comprehension make
independent contributions to reading comprehension, as found in prior research,4 a
general linear modelling (GLM) forward-fitting approach was adopted to model the
effects of age, non-verbal IQ, group (RE and Controls), word reading ability
(composite), and listening comprehension, on the YARC reading comprehension
score. The Bayesian Information Criterion Index was used to ascertain the best-fitting
model. All assumptions of this test were met and the residuals were checked on the
model of best fit (reported in Table II).
Several clinical variables were examined in relation to the study measures,
using raw scores. Age of onset was analysed using partial correlations controlling for
age. Separate ANCOVAs of sub-groups were conducted for medication status (AED,
none, control) with age as a covariate. Planned analyses of sub-groups based on
seizure frequency and occurrence of SSD were not conducted due to low numbers
(see Table II). Partial correlations controlling for age examined whether DCD scores
correlated with any reading measures. G-F articulation measures were assessed by
ANCOVA (group differences) and partial correlation (relation to reading and
comprehension measures).
RESULTS
Differences between rolandic epilepsy and control groups
Figure 2 provides the mean differences and 95% confidence intervals (z-
scores) for each literacy measure. Single factor ANCOVAs with group (RE, Control)
and age entered as a covariate were conducted on the raw ability scores. Reading
comprehension and word reading composite scores were significantly lower in RE
children than controls, F(1, 61) = 6.89, p = .01, p2
= .10, and F(1, 61) = 6.84, p = .01,
p2
= .10, respectively. Although children with RE performed worse on the test of
listening comprehension this was marginal F(1, 61) = 3.81, p = .06, p2
= .06. Non-
verbal IQ was significantly worse in the RE sample, F(1, 61) = 10.53, p = .002, p2
=
.15.
Table I reports the number of children in the RE and control samples with
significant impairment (one standard deviation below standard score mean) on each
measure (excluding non-verbal IQ and listening comprehension). A significantly
greater number of children in the RE sample fell into this category for YARC reading
accuracy (p = .02, Fisher’s exact test) and TOWRE sight word reading (p = .04), and
marginally more for YARC reading comprehension (p = .06). The p values on all
other measures were greater than p = .19.
Relation between word-reading ability, listening comprehension and reading
comprehension
The effects of age, non-verbal IQ, group (RE and Controls), word reading
(composite), and listening comprehension, on predicting the reading comprehension
score were examined. Table II shows the model of best fit (accounting for 70% of
variance in reading comprehension scores), which included significant main effects of
word reading and listening comprehension, demonstrating that each make significant
independent contributions to reading comprehension, to similar degrees in both the
RE and control group.
7
Effects of clinical variables on reading comprehension, word-reading and
listening comprehension.
Single factor ANCOVAs examining the effect of medication status (RE no
medication, RE medication, controls) with age as a covariate, were conducted. Table
SI reports the means and SDs (in standard scores) for the reading and comprehension
measures as a function of medication status. For reading comprehension, children in
the RE medication group were poorer than controls: F(2, 60) = 6.00, p = .004, p2
=
.17; mean difference = -10.44, 95% CI -17.87-(-3.02). The same pattern was found
for word reading (composite): F(2, 60) = 5.67, p = .006, p2
= .16; mean difference = -
12.93, 95% CI -22.39-(-3.47), and also non-verbal IQ: F(2, 60) = 8.13, p = .001, p2
=
.21; mean difference = -5.66, 95% CI -9.12-(-2.20). In all comparisons, RE children
not in receipt of AEDs did not differ significantly from the other two groups in the
post hoc tests. There was no effect of group on listening comprehension, F(2, 61) =
2.26, p = .11, p2
= .07. There were no significant (partial) correlations between age
of onset and any study measures.
Are SSD or DCD more likely to be associated with reading difficulties in children
with RE than in controls?
The occurrence of SSD was too low (RE: n = 4, TD: n = 1) to permit analysis.
However, children with RE were more likely to have articulation difficulties than
controls. A single factor ANCOVA on the raw G-F test scores with group (RE,
control) as a factor and age as a covariate revealed a significant effect of group, F(1,
62) = 5.19, p = .03, p2
= .08, although all children scored within age appropriate
range (all scores > 85). There were no significant correlations between the G-F test
scores and the literacy measures (all rs < -.32, all ps > .05). Of the four RE children
with a history of SSD, two had significant word reading problems (more than one
standard deviation below the standardised score mean).
For children with RE there was a significant correlation between the overall
raw score on the DCD questionnaire and the word reading composite, r (22) = .52, p
= .009, CI .09 - .78. Correlations with all other literacy measures for both groups were
small (all rs < -.35, all ps > .06).
DISCUSSION
In line with predictions, children with RE were significantly poorer than
controls on measures of word reading, reading comprehension and non-verbal IQ.
The RE group was heterogeneous; not all had poor reading comprehension and, where
apparent, there was evidence that neither poor word reading nor listening
comprehension was the consistent predictor. Our results support a recent meta-
analysis highlighting word reading ability as a potential weakness in children with
RE1. However, our study identifies reading comprehension as an additional area of
vulnerability.
Word reading and listening comprehension made independent contributions to
reading comprehension performance for both groups. The contributions of these
variables were in addition to the variance explained by IQ, consistent with other work
on reading in non RE populations20, 21
. Recent work reports differences between
children with RE and controls on language measures regardless of IQ, although
differences were larger when IQ was low.1 Taken together these findings support the
argument that IQ is not the main independent cause of the word reading and reading
comprehension difficulties children with RE may experience. Future work examining
8
the relation between a broader range of receptive language and cognitive skills, such
as working memory and executive skills, on reading comprehension outcomes is
required to establish the critical pressure points in the reading system for children
with RE22
.
Although the occurrence of SSD was low, there was some evidence of residual
articulation difficulties for children with RE. Our study supports the proposal that
word reading difficulties in children with RE may be associated with speech or more
generalised dyspraxic difficulties.3, 11, 12, 23
However, neither SSD nor DCD was
evident in all children with RE who had reading difficulties. Whether dyspraxia is co-
occurring or causally related to the reading difficulties associated with RE, at least for
some children, is a question for future research.
In terms of clinical variables, the lack of relation between our reading and
comprehension measures and age of onset is consistent with other studies 8, 24
. Our
findings of worse performance on reading measures relative to controls for the RE
group on medication is consistent with the findings of Ay and colleagues25
. However,
in our sample, it is quite possible that medication status was confounded with a more
complex epilepsy, so causal inferences cannot be drawn.
Limitations of the study include the use of only a single indicator for each of
two key variables, reading and listening comprehension, resulting in a somewhat
narrow assessment of these complex constructs. Additionally, our sample was smaller
than expected, due to the difference between diagnostic decisions occurring at
secondary (general paediatric) versus tertiary (paediatric neurology) levels, potentially
resulting in a loss of power for reliably detecting differences in listening
comprehension between RE and control groups. Our sample size was in line with
other recent cross-sectional studies of children with RE 1, however.
To summarise, in this first study examining the full reading profiles of
children with RE, we found that reading comprehension was compromised, in
addition to previous reported difficulties with word reading. Critically, we
demonstrated that reading comprehension was determined by both word reading and
listening comprehension. Despite the limitations noted, educators and clinicians
working with children with RE should consider these children to be at risk of word
reading and listening comprehension difficulties. Based upon the profile of difficulties
presented by individual children with RE, intervention for one or both of these aspects
of reading may be required, to support the ultimate goal of reading comprehension.
9
ACKNOWLEDGEMENTS
This research was funded by the Waterloo Foundation (grant reference: 1134-1593).
The authors would like to thank the parents, children and schools for their
participation in the research. The authors would also like to thank the Medicines for
Children Research Network, hospital site PIs and research nurses for their support
during recruitment and Samantha Seed, Emma James and Chiara Anceschi for their
assistance during data collection and scoring. We would also like to acknowledge
Professor Damon Berridge (Swansea University) for his work on the power
calculations at the outset of this research.
10
REFERENCES
1. Smith AB, Bajomo O, Pal DK. A meta‐analysis of literacy and language in children with
rolandic epilepsy. Developmental Medicine & Child Neurology 2015; 57:1019-26.
2. Shinnar S, O’Dell C, Berg AT. Distribution of epilepsy syndromes in a cohort of children
prospectivley monitored from the time of their first unprovoked seizure. Epilepsia
1999; 40: 1378-83.
3. Clarke T, Strug LJ, Murphy PL, et al. High risk of reading disability and speech sound
disorder in rolandic epilepsy families: case–control study. Epilepsia 2007; 48: 2258-
65.
4. Language and Reading Research Consortium. Learning to read: should we keep things
simple? Reading Research Quarterly 2015; 50:151-69.
5. Darcovich N, Jones S, Kirsch IS, et al. Literacy skills for the knowledge society: further
results from the International Adult Literacy Survey. Paris, France: OECD, 1997.
6. Perfetti C, Landi N, Oakhill J. The acquisition of reading comprehension skill. In:
Snowling MJ, Hulme C, editors. The science of reading : a handbook. Oxford:
Blackwell, 2005.
7. Cain K, Oakhill J, Bryant P. Investigating the causes of reading comprehension failure: the
comprehension-age match design. Reading and Writing: An Interdisciplinary Journal
2000; 12: 31- 40.
8. Overvliet GM, Besseling RMH, van der Kruijs SJM, et al. Clinical evaluation of language
fundamentals in Rolandic epilepsy, an assessment with CELF-4. European Journal of
Paediatric Neurology 2013; 17: 390-396.
9. Croona C, Kihlgren M, Lundberg S, Eeg-Olofsson O, Eeg-Olofsson KE.
Neuropsychological findings in children with benign childhood epilepsy with
centrotemporal spikes. Developmental Medicine and Child Neurology 1999; 41: 813-
818.
10. Tunick R, Pennington B. The etiological relationship between reading disability and
phonological disorder. Annals of Dyslexia 2002; 52: 75-97.
11. Peter B, Raskind W. Evidence for a familial speech sound disorder subtype in a
multigenerational study of oral and hand motor sequencing ability. Topics in
Language Disorders 2011; 31:145-67.
12. White S, Milne E, Rosen S, et al. The role of sensorimotor impairments in dyslexia: a
multiple case study of dyslexic children. Developmental Science 2006; 9: 237-55.
13. Snowling MJ, Stothard S, Clarke P et al. York assessment of reading for comprehension.
London: GL Assessment, 2009.
14. Colenbrander D. Understanding the role of oral vocabulary in reading comprehension
difficulties. Unpublished PhD Thesis. Sydney, Australia: Macquarie University, 2014.
15. Torgesen J, Wagner R, Rashotte C. Test of word reading efficiency second edition
(TOWRE-2). Austin, Texas: PRO-ED, 2012.
16. Semel E, Wiig E, Secord W. Clinical evaluation of language fundamentals fourth UK
edition (CELF 4-UK). London, UK: Harcourt Assessment, 2006.
17. Wechsler D. Wechsler intelligence scale for children - Fourth UK Ed (WISC-IV UK).
London: Pearson, 2003.
18. Goldman R, Fristoe M. Goldman-Fristoe test of articulation 2. Minneapolis, USA:
Pearson, 2000.
19. Wilson B, Crawford S. The developmental coordination disorder questionnaire
(DCDQ'07) 2007 [cited July 2016]. Available from: http://www.dcdq.ca/.
20. Oakhill JV, Cain K. The precursors of reading ability in young readers: evidence from a
four-year longitudinal study. Scientific Studies of Reading 2012; 16: 91-121.
11
21. Catts HW, Adlof SM, Weismer SE. Language deficits in poor comprehenders: a case for
the simple view of reading. Journal of Speech, Language, and Hearing Research
2006; 49: 278-93.
22. Language and Reading Research Consortium. Pressure points in reading comprehension:
A quantile multiple regression analysis. Journal of Educational Psychology in press.
23. Pal DK, Li W, Clarke T, Lieberman P, Strug LJ. Pleiotropic effects of the 11p13 locus
on developmental verbal dyspraxia and EEG centrotemporal sharp waves. Genes,
Brain and Behavior 2010; 9:1004-12.
24. Tedrus GMAS, Fonseca LC, Melo EMV, Ximenes VL. Educational problems related to
quantitative EEG changes in benign childhood epilepsy with centrotemporal spikes.
Epilepsy & Behaviour 2009; 15: 486-90.
25. Ay Y, Gokben, S, Serdaroglu, G et al. Neuropsychologic impairment in children with
rolandic epilepsy. Pediatric Neurology 2009; 41: 359-63.
12
Figure 1. Participant recruitment for the children with rolandic epilepsy.
Identified by local
site and gave
consent for research
team to contact
n=40
Failed telephone
contact (n=7)
Declined to participate
(n=2)
Diagnosis changed to
alternative form of
epilepsy (n=1)
Total recruited
n = 30
(75%) Declined to participate
prior to testing (n=1)
Excluded due to
insufficient evidence
for diagnosis (n=4)
Final sample
n = 25
13
Table I. Participant characteristics and study measures expressed as standardised
scores
Rolandic epilepsy
(N = 25)
Control
(N = 39)
Gender Male/Female 16/9 25/14
Mean age, y:mo (SD) 9:1 (1:7) 9:1 (1:3)
History of speech/language problems a
Received Speech Therapy
Indication of Speech Sound Disorder
Indication of possible Developmental
Coordination Disorder
7 (29.2%)
6 (25.0%)
4 (16.7%)
13 (54.2% )
3 (10.0%)
3 (10.0%)
1 (3.3%)
3 (9.7%)
Clinical Variables n (%)
Age at onset, y:mo (SD) 7:0 (1:10)
Antiepileptic medication
None
Monotherapy
Polytherapy
12 (48.0%)
12 (48.0%)
1 (4.0%)
Laterality (n)b
Right
Left
9 (36.0%)
5 (20.0%)
Bilateral
Evolving
11 (44.0%)
6 (24.0%, 2 left, 1
right, 3 bilateral)
Seizure type
With motor/autonomic components
Sensory/psychic phenomena
Evolving to bilateral convulsive
seizure
25 (100.0%)
10 (40.0%)
13 (52.0%, 4 with
sensory/psychic
phenomena)
Atypical (n)
Status
Epilepticus
Diurnal
seizures
Screaming
1 (4.0%)
8 (32.0%)
2 (8.0%)
Todds Paresis
Leg jerking /lateral
body torsion
Epigastric Pain
4 (16.0%)
7 (28.0%)
1 (4.0%)
Seizure Frequency (n)
Less than monthly
Monthly
Weekly
21 (84.0%)
3 (12.0%)
1 (4.0%)
Comorbid diagnosis (n)c
None
Present
19 (76.0%)
6 (24.0%)
Standardised Measures
Mean (SD) Mean (SD)
Reading Comprehension (YARC)
Significant impairment d
Reading Accuracy (YARC)
Significant impairment
Reading Rate (YARC)e
Significant impairment
Sight Word Reading (TOWRE)
Significant impairment
Phoneme Decoding (TOWRE)
100.84 (12.79)
3 (12.0%)
98.12 (12.84)
4 (16.0%)
95.52 (15.25)
5 (20.0%)
94.32 (18.94)
7 (28.0%)
95.44 (18.57)
108.23 (11.12)
0
103.62 (10.80)
0
105. 03 (15.01)
4 (10.26%)
103.64 (14.07)
3 (7.69%)
101.69 (11.78)
14
Significant impairment
Listening Comprehension (CELF)f
Nonverbal IQf
Articulation (Goldman Fristoe)
Significant impairment
7 (28.0%)
9.16 (3.16)
7.60 (2.90%)
103.16 (4.02)
0
5 (19.23%)
10.62 (2.71)
10.08 (2.77)
104.36 (2.78)
0
Note. For all tests M = 100 and SD = 15 with the exception of the Listening
Comprehension (CELF) and non-verbal IQ (WISC) where M = 10 and SD =3. a
One
child with RE and nine control children had non-returned language questionnaires.
One child with RE and eight control children had non-returned DCD questionnaires. b
In 80% of cases laterality was fully confirmed by EEG. c Comorbid
diagnoses: 1
Anxiety, 1 ADHD and resolved hearing difficulty, 2 ASD and 2 movement disorder. d
Number of children one standard deviation or below the standard score mean. e Four
children had missing data for reading rate. For these children the word reading
composite was calculated using the three available scores. f Only one subtest used
therefore percentage with significant impairment not reported.
15
Figure 2.
Mean difference between RE and control group (95% CIs) for raw scores (expressed
as z-scores) on the study measures.
Reading comprehension (YARC)
Reading accuracy (YARC)
Reading rate (YARC)
Single word reading (TOWRE)
Single non-word reading (TOWRE)
Listening comprehension (CELF)
Word reading composite
Non-verbal IQ (WISC)
-1.3 -1.1 -0.9 -0.7 -0.5 -0.3 -0.1 0.1 0.3
16
Table II. Parameter estimates for best-fit general linear model of reading
comprehension, with word reading and listening comprehension as predictors.
Parameter (z-transformed) Estimate
()
SE 95% CI p R2
Word-reading .65 .07 .51-.79 <.0001 .70
Listening Comprehension
.39 .07 .25-.53 <.0001
Note. The model including NVIQ was a similar fit (BIC (word reading, listening
comprehension) = 120.02, BIC (word reading, listening comprehension, NVIQ) =
120.41) and accounted for a similar proportion of variance in reading comprehension
(.72), thus the more parsimonious model is preferred. The residuals did not depart
from normality, there was no difference between groups in residual variance and the
residuals were uncorrelated, Durbin-Watson = 2.29.